19:05:41 <decentral1se> #startmeeting Ansible Molecule Working Group
19:05:41 <zodbot> Meeting started Wed May 15 19:05:41 2019 UTC.
19:05:41 <zodbot> This meeting is logged and archived in a public location.
19:05:41 <zodbot> The chair is decentral1se. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:05:41 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:05:41 <zodbot> The meeting name has been set to 'ansible_molecule_working_group'
19:05:48 <decentral1se> OK, who's around?
19:06:16 <fabianvf> I'm around
19:06:26 <decentral1se> #chair fabianvf
19:06:26 <zodbot> Current chairs: decentral1se fabianvf
19:08:58 <fabianvf> cool this might be a quick one it seems
19:09:45 <decentral1se> #info Agenda https://github.com/ansible/community/issues/427
19:09:52 <decentral1se> sure thing ;)
19:11:32 <fabianvf> So the documentation for the ansible verifier is merged and viewable on the master version of the docs site
19:11:41 <fabianvf> https://molecule.readthedocs.io/en/master/configuration.html#id22
19:11:54 <decentral1se> #topic New Ansible verifier
19:12:25 <decentral1se> that's great
19:12:33 <decentral1se> #info https://github.com/ansible/molecule/issues/2013#issuecomment-488392662
19:12:39 <decentral1se> have you seen this comment from latest WG?
19:12:51 <decentral1se> We discussed that we need to further develop the documentation
19:13:05 <fabianvf> ah yeah, just the basic docs are up right now
19:13:56 <fabianvf> ok that all makes sense
19:14:05 <fabianvf> do we have a place in documentation for that to go currently?
19:14:16 <decentral1se> good question, unsure
19:14:31 <fabianvf> or is there a docs.ansible.com source with info we could link to?
19:14:48 <fabianvf> trying to figure out if we do anything like this for the other verifiers currently
19:14:54 <decentral1se> From what I remember, the only link we had to some existing examples is on that ticket
19:15:54 <decentral1se> Could go on https://molecule.readthedocs.io/en/stable/examples.html?
19:17:16 <fabianvf> yeah, there's also this page: https://docs.ansible.com/ansible/latest/reference_appendices/test_strategies.html
19:18:00 <decentral1se> Oh that is good
19:18:09 <decentral1se> some parts are out of context, but it is nice
19:18:13 <fabianvf> perhaps we should look at updating the basic ansible docs as well now that molecule provides a solid testing framework
19:20:10 <decentral1se> That's a nice idea
19:20:29 <decentral1se> probably the easiest right now is to link to that doc and whatever else in the Examples section of molecule docs
19:21:21 <fabianvf> yeah, we should probably link there in the configuration section as well
19:21:28 <decentral1se> then when 2.8 is out and we make a next molecule release, we can work on upstream patches etc.
19:21:44 <decentral1se> yep
19:21:52 <fabianvf> the other verifiers link to the official docs from the configuration section
19:21:59 <decentral1se> ok cool
19:22:15 <decentral1se> I'll update the ticket with these ideas
19:23:45 <decentral1se> https://github.com/ansible/molecule/issues/2013#issuecomment-492790060
19:24:10 <decentral1se> Hopefully someone will have some cycles to pick this up soon
19:24:57 <fabianvf> cool
19:25:39 <decentral1se> ok, other thing I had
19:25:50 <decentral1se> #topic molecule release schedule
19:25:51 <fabianvf> if there's any sort of larger docs overhaul to coincide with collections and stuff we should try to get that testing_strategies page included
19:26:16 <decentral1se> yeah, looking forward to see how that goes down
19:26:33 <decentral1se> think jtanner has something cooking for that. We were due a planning session of some sort
19:26:40 <decentral1se> will hopefully happen some WG soonish
19:26:58 <decentral1se> sorry, rushed the topic switch, something further, please shoot
19:27:06 <fabianvf> nope that was my last idea
19:27:14 <fabianvf> on to releases!
19:27:23 <decentral1se> yeah right, so, current status as I understand it is
19:27:24 <jtanner> i forked molecule to jctanner/ansible-sdk to try to figure out some things releated to collections
19:27:32 <fabianvf> since we're the only ones here I guess we get to set the release schedule now :P
19:28:00 <decentral1se> hehe
19:28:11 <decentral1se> jtanner: ah right, any progress report? ;)
19:28:40 <jtanner> no, still working on prototyping standalone ansible-test to integrate with it
19:28:50 <jtanner> and what all that means for a collection
19:29:17 <decentral1se> as in, ansible-test will invoke molecule or?
19:29:27 <jtanner> molecule -> ansible-test -> sanity
19:29:57 <jtanner> trying to bring all the stuff we do upstream with plugins down to the molecule world
19:30:18 <jtanner> since that is 99% of what collections are about
19:30:49 <decentral1se> ok, nice to know
19:31:01 <decentral1se> RE: schedule
19:31:03 <decentral1se> #info https://github.com/ansible/molecule/issues/2042
19:31:24 <decentral1se> CI is blocked and this inspec test is repeatedly failing (nobody can reproduce locally so far)
19:31:49 <decentral1se> we have a critical bug which stops molecule running at all regarding yamllint (AFAICT)
19:32:11 <decentral1se> #info https://github.com/ansible/molecule/issues/2012
19:32:34 <decentral1se> Latter has a fix out but is blocked by the former :)
19:33:20 <decentral1se> After we would unstick these two, I think we can then nag webknjaz to make a pre-release
19:35:05 <fabianvf> awesome
19:35:19 <decentral1se> We know have a +1, +2 review policy in place to stop rash PRs being dropped (hehe)
19:35:26 <decentral1se> think that was some part of the code churn
19:36:07 <decentral1se> but otherwise, hoping we can focus on stabilising, getting Ansible verifier out and just fixing bugs against that from the pre-release
19:36:23 <decentral1se> Whoever has cycles, please join in and things will land sooner :)
19:36:45 <fabianvf> I'm hoping in the next week or so people will come out of the post-summit fugue
19:37:01 <decentral1se> oh, a summit?
19:37:28 <fabianvf> Red Hat summit was last week
19:37:48 <decentral1se> aha, right
19:38:17 <fabianvf> lots of demos needed to be prepped, product releases finished and put through the downstream release processes etc
19:38:27 <decentral1se> busy times
19:38:44 <decentral1se> that's all I had on the schedule stuff
19:38:53 <decentral1se> any thoughts?
19:39:11 <fabianvf> I think that's good, we can hold discussion on what the broader release cycle should look like for when we have more participants
19:39:36 <decentral1se> agreed
19:39:55 <decentral1se> ok, open floor then
19:40:14 <decentral1se> #chairs jtanner
19:40:27 <jtanner> i have nothing
19:40:31 <fabianvf> do we have a documented process for submitting new platforms?
19:40:41 <decentral1se> #chair jtanner
19:40:41 <zodbot> Current chairs: decentral1se fabianvf jtanner
19:40:49 <decentral1se> ah, that was it ... forgot to run it ;)
19:41:12 <fabianvf> I was looking at writing one for the official `kind` tool which is great for spinning up local kubernetes clusters
19:41:24 <fabianvf> `kind` = Kubernetes-in-Docker
19:41:53 <decentral1se> Ah, I remember you mentioning this earlier in the agenda
19:42:39 <decentral1se> "new platforms" do you mean "new drivers"?
19:42:44 <fabianvf> drivers, sorry
19:42:48 <decentral1se> ah right
19:42:58 <decentral1se> well, no, then I can confirm there is no documented process
19:43:10 <decentral1se> I've blundered my way through the Linode one and now the Hetzner Cloud one too
19:43:19 <decentral1se> you could check those PRs ... it's not pretty
19:43:33 <fabianvf> cool, that's what I figured, was the same writing the Ansible verifier
19:43:42 <decentral1se> yep
19:43:51 <decentral1se> is `kind` a "local only" thing?
19:43:54 <fabianvf> just kind of hard to find the exact API you need to expose and then especially the tests
19:44:01 <fabianvf> yeah
19:44:27 <decentral1se> So we wouldn't have "the k8s driver" but a test driver or something
19:44:49 <fabianvf> right, it would let you "mock out" real kubernetes infra
19:44:54 <decentral1se> There were some grumblings about adding the Hetzner Cloud driver (yet a new driver, maintainence issues)
19:45:10 <fabianvf> yeah a plugin system seems ideal
19:45:44 <fabianvf> though this one would at least be testable without credentials or anything
19:46:26 <decentral1se> Is it well adopted by the k8s community? I mean, would people find that normal to use
19:46:38 <decentral1se> and then naturally drop into k8s in their main task definitions outside of molecule
19:47:36 <fabianvf> yeah, it's official: https://github.com/kubernetes-sigs/kind
19:47:51 <fabianvf> I'm not sure about the second part of your question though
19:47:58 <fabianvf> what do you mean?
19:48:12 <decentral1se> ah cool
19:48:59 <decentral1se> rephrasing: would it make sense for a role that uses standard k8s modules
19:49:07 <decentral1se> to use a molecule kind driver to test it
19:49:26 <fabianvf> yeah I think so, I'm doing it in a less-official way right now in Ansible Operator
19:49:31 <decentral1se> in terms of: kind can reproduce the type of environment that the standard k8s modules create
19:49:42 <fabianvf> https://github.com/fabianvf/game-server-operator/blob/master/molecule/default/molecule.yml
19:50:03 <fabianvf> yeah, kind is a real kubernetes cluster
19:50:08 <fabianvf> or creates one rather
19:50:17 <decentral1se> interesting
19:50:29 <decentral1se> sounds well worth taking on then
19:50:36 <fabianvf> https://github.com/fabianvf/game-server-operator/blob/master/molecule/default/asserts.yml
19:50:53 <decentral1se> # topic `kind` kubernetes driver
19:50:56 <decentral1se> #topic `kind` kubernetes driver
19:50:58 <decentral1se> dammit
19:51:09 <fabianvf> we're using molecule to do full e2e integration tests of operators
19:51:57 <decentral1se> oh this is cool
19:52:48 <decentral1se> I like that there is no tests/ folder and you just have verify.yml
19:52:58 <decentral1se> or asserts.yml I mean
19:53:14 <fabianvf> it will be verify.yml, that code predates the ansible verifier
19:53:47 <fabianvf> but molecule is so nice for working with kubernetes, I've started using it for e2e tests of non-ansible projects at well
19:54:00 <fabianvf> s/at/as/
19:54:33 <decentral1se> very nice
19:54:42 <decentral1se> I would recommend then to open an issue about the driver then, no?
19:55:00 <fabianvf> yeah, I should do that
19:55:15 <fabianvf> usually I go straight to opening a PR :P
19:55:48 <fabianvf> I'll post the issue today
19:55:58 <decentral1se> I usually do the same but my spider sense tingles when I do that on this project
19:56:10 <decentral1se> cool
19:56:19 <fabianvf> would this be a feature request or a proposal?
19:56:30 <decentral1se> feature request, I guess! Nothing majorly new, just a new thing
19:56:38 <fabianvf> cool
19:56:57 <decentral1se> right, I'm gonna drop off if there is nothing else
19:57:01 <fabianvf> I don't have anything else to bring up
19:57:09 <decentral1se> solid stuff
19:57:15 <decentral1se> thanks to all then :)
19:57:44 <decentral1se> let's see if the party starts again when the summit hangover ends ...
19:58:00 <decentral1se> #endmeeting